Facility Management KPIs
A facilities scorecard where every measure carries how it is calculated, written out in words. Most KPI lists give you names, and two sites then report different numbers for the same measure without either of them being wrong.
- Twenty measures, each with its calculation written in words
- RAG thresholds agreed once, before the first report
- An action required only on the lines that are Amber or Red
- How every target was set, recorded beside the target
Facility Management KPIs
Named, and defined
| # | KPI DEFINITIONS | SOURCE | FREQ | OWNER |
|---|---|---|---|---|
| 1 | Statutory compliance, completed on time over tasks due | |||
| 2 | First-time fix rate, and what counts as a first-time fix | |||
| 3 | Reactive to planned ratio, by job count not by hours |
The document you will get. Download for the full, editable file.
Who this FM KPI scorecard is for
One scorecard, and three audiences who each need something different from the same page.
Whoever compiles the report
The definitions are yours. Once the calculation is written down, compiling stops being a judgment call every month and the number stops moving because a different person produced it.
The client or executive reading it
You get a measure you can interrogate. Being able to ask what counts as a first-time fix, and get the same answer twice, is what makes a scorecard worth reading.
The site or contract manager
The RAG and the action list. Only the Amber and Red lines carry an action, which is what keeps the report a working document rather than a wall of commitments.
Whoever compares sites
Definitions are what make comparison legitimate. Two sites reporting the same KPI on different rules produce a league table that measures the rules, not the sites.
Where the definitions matter most
The measures are broadly common. What changes is which ones get asked about and how much rides on them. If you run one of these, the sector page goes further than the template does.
- FM service providers
Contracted reporting, where the definition is effectively contractual and a loose one is argued about at every review.
FM service provider software - Commercial real estate
Multi site portfolios, where the whole point is comparison and comparison only works on identical rules.
Portfolio maintenance software - Healthcare
Statutory compliance dominates, and the definition decides whether a deferred test counts as done or due.
Healthcare maintenance software - Corporate facilities
Occupant satisfaction and space measures carry as much weight as plant, and both are easy to define loosely.
Corporate facilities software - Retail and malls
Many small sites and thin data, where a measure is only as good as the coverage behind it.
Retail maintenance software - Education
Seasonal occupancy, which makes any per-occupant measure meaningless unless the basis is stated.
Campus maintenance software
What a facility management KPI set should contain
A facility management KPI set is the list of measures a facilities operation reports on, with the calculation for each one written out rather than assumed. It records what each measure tells you, how it is worked out, where the number comes from, who owns it and how often it is reported, then tracks target against actual.
A. The sections that carry the weight
| Field | What goes in it | Why it earns its place |
|---|---|---|
| Scorecard details | The site or portfolio, the period and who prepared it | A scorecard with no stated period and no author becomes impossible to compare with the one before it. |
| The measure, and what it tells you | A plain sentence on what the number is for | Worth writing because it exposes measures nobody can justify. A KPI whose purpose cannot be stated in a sentence is usually inherited rather than chosen. |
| How it is calculated | The formula in words, including what counts and what does not | The reason this document exists. Everything else here is ordinary; this column is what stops two sites reporting different numbers for the same measure. |
| Source | Where the number actually comes from | A measure with no named source is an estimate. Naming it also reveals which measures depend on somebody typing them in by hand. |
| Frequency and owner | How often it is reported and who is accountable for it | An unowned measure drifts. Frequency matters too, because a monthly measure compared against a quarterly one is not a comparison. |
| Target, actual and variance | What was expected, what happened and the gap, calculated | Variance works itself out, which matters because a hand-typed variance is the easiest number in the file to get quietly wrong. |
| RAG status | Green, amber or red against agreed thresholds | Spelled exactly, because the summary counts it. The thresholds belong in the file rather than in the head of whoever is presenting. |
| Trend | Whether this measure is improving, flat or worsening | Direction often matters more than level. A red measure improving for three periods is a better position than an amber one sliding. |
| Scorecard status | How many measures are green, amber and red | The one-line summary an executive reads first, and the reason RAG has to be spelled consistently. |
| Thresholds | What green, amber and red actually mean, in numbers | Set once, before the first report. Deciding the band after seeing the result is the most common way a scorecard becomes decorative. |
| Actions from this period | Only for the measures that are amber or red | What turns a report into work. Restricting it to the lines that are off target is what keeps the list short enough to finish. |
| Choosing and changing measures | Why these, what you deliberately do not measure, and how targets were set | The section that ages best. It answers the question every new reader asks, which is why this set and not some other one. |
Write the calculation out in words, including what does not count, because the exclusions are where two honest people diverge. First-time fix rate is the classic example. One site counts a job as fixed first time if it was closed on the first visit, which quietly includes every job closed because nobody could get access. Another excludes no-access closures entirely. Both are defensible, both are called first-time fix rate, and the two numbers can differ by ten points or more on the same operation. Nothing about that is visible on a scorecard that just names the measure. The same trap sits under statutory compliance, where the question is whether a deferred task counts as done, due or excluded, and under backlog, where it is whether the clock starts at the raise date or the due date. None of these need a right answer. They need a written one.
B. What it looks like filled in
Four measures as the file defines them. The middle column is the one every KPI list has, and the right-hand one is the one that makes the number reproducible.
| KPI | What it tells you | How it is calculated |
|---|---|---|
| Statutory compliance | Whether legally required tasks are being done | Statutory tasks completed on time, over statutory tasks due |
| First-time fix rate | Whether a job is resolved on the first visit | Jobs closed on the first visit, over jobs closed, excluding no-access |
| Reactive to planned ratio | Whether the program is running or reacting | Reactive jobs over planned jobs, counted by job not by hour |
| Backlog hours | Work committed to but not yet done | Estimated hours on open jobs past their due date |
Read the second row twice, because the last three words are doing more work than the rest of the scorecard. Excluding no-access closures from first-time fix rate is a choice, and the opposite choice is equally defensible. What is not defensible is leaving it unstated, because then the measure reports on whoever compiled it. The same applies to the third row: reactive to planned counted by job treats a ten minute reset and a two day repair as equal, while counting by hour tells a different story about the same month. Neither is wrong. A scorecard that does not say which one it used is. This is also why the fourth row specifies that backlog starts at the due date rather than the raise date, which is the difference between measuring work that is late and measuring work that merely exists.
Word for the version that goes into a contract or a reporting pack, Excel for the one that calculates variance and counts the RAG statuses, with a Trend tab holding a row per period, and PDF for circulation. Fill the definitions once and the scorecard becomes routine. Free, and yours to rebrand.
How do you report facility management KPIs?
Do the definitions once, properly, before the first report. Everything after that is filling in two columns a month. Six steps.
Choose the measures, and write down why
Pick the set deliberately rather than inheriting it. The file asks what you deliberately do not measure as well, because an unstated exclusion is the thing people later assume is covered.
Write the calculation for each one in words
Include what counts and what does not. This is the step that takes an afternoon and saves an argument a month, and it is the step almost everybody skips.
Name the source, the frequency and the owner for each
Where the number comes from, how often it is reported and who is accountable. A measure with no named source is an estimate that will be presented as a fact.
Set the targets, and record how you set them
Benchmark, contract, previous performance or aspiration. A target lifted from another portfolio's benchmark will be missed all year and will teach people to ignore the whole report.
Agree the RAG thresholds before the first report
Put the numbers in the file. Deciding what counts as amber after seeing the result is how a scorecard turns into a presentation.
Report the period, then raise actions only on amber and red
Actual against target, variance, RAG and trend. Then an action, an owner and a date on the lines that are off target, and nothing on the lines that are not.
A measure with a name versus a measure with a formula
The same measure on two scorecards. One of them can be reproduced by somebody who was not in the room, and the difference is one column.
| Aspect | A measure with a formula | A measure with only a name |
|---|---|---|
| Two sites reporting it | Produce the same number | Produce two numbers, both honest |
| What a reader can ask | Whether the rule is the right one | Only whether they believe the figure |
| When the compiler changes | The number stays comparable | The number moves for no reason |
| Comparison across a portfolio | Legitimate | Measures the rules, not the sites |
| What an exclusion does | Is stated, and can be argued | Hides inside the number |
| Cost of getting it wrong | An hour agreeing the wording | A year of numbers nobody trusts |
The failure here is quiet, which is why it survives for years. Nothing looks wrong on a scorecard whose measures are only named. The numbers are plausible, the trend lines move, and the report gets presented every month. It only surfaces when two sites are put side by side and one is apparently far better, at which point the argument is about definitions rather than performance and everybody discovers there were never any. Writing the formula is not a governance nicety. It is what makes the difference between a number that describes an operation and a number that describes whoever produced it.
When the template starts to feel limiting
The file is built to be defined once and filled in each period. Four things start to hurt as soon as you want it to be live.
Every number is still typed in
The definitions make the measure reproducible; they do not produce it. Somebody still collects each figure by hand, which is where most of the effort and all of the errors are.
The scorecard is always looking backward
By the time a period is compiled, the month it describes is over. Nothing here can tell you a measure is sliding while there is still time to act on it.
Coverage is invisible in the number
A measure taken from a source covering half the estate looks exactly like one covering all of it. The file asks for known limitations for that reason, and nobody reads them.
Comparing periods means opening files
The Trend tab helps for one site. Asking which of eleven sites has been red on statutory compliance twice running means opening eleven workbooks.
What running this in Facilio looks like
The measures do not change, and the definitions still matter. What changes is that the number is computed from the work instead of collected from people.
Ops Performance Intelligence
One rule, applied everywhere
Each measure is computed from the underlying jobs on a single definition, so two sites cannot quietly count differently and a portfolio comparison means something.
Work Order Intelligence
The figures are current, not compiled
Compliance, response and backlog come from the work as it closes, so a measure sliding mid-period is visible while there is still time to do something.
Audit Report Intelligence
The number can be opened
Every figure traces back to the jobs behind it, so a challenged measure is answered by showing the records rather than by re-explaining the spreadsheet.
Asset Intelligence
Coverage stops being a footnote
Because measures are built from records rather than from a submitted sheet, a gap in coverage shows up as missing data instead of a confident average over whatever happened to be counted.
Hallucination-free by design. Atom AI answers from the records in your tenant rather than generating plausible text, so an empty field reads as empty rather than filled in for you.
Frequently asked questions
What are facility management KPIs?
They are the measures a facilities operation reports on: compliance, response and resolution, cost, energy, space, satisfaction and contractor performance. This template carries twenty of them.
What distinguishes this one from a list of KPI names is that every measure carries its calculation written out in words, along with where the number comes from, how often it is reported and who owns it.
How many KPIs should a facilities team report?
Fewer than most teams start with. The file defines twenty so you can choose from a considered set, and a working scorecard is usually ten to twelve measures with a genuine owner each.
The test is whether anybody acts on a measure. A KPI that has never changed a decision is costing collection effort every period and is a candidate for the deliberately-not-measured list.
Why write out how each KPI is calculated?
Because the same measure name routinely produces different numbers. First-time fix rate that includes jobs closed for no access is a different figure from one that excludes them, often by a wide margin, and both are called first-time fix rate.
Writing the rule down, including what does not count, is what makes the measure reproducible by somebody who was not there and what makes comparison between sites legitimate.
What is the difference between this and a facility management plan?
This is about definition and reporting: what is measured, how it is calculated and how it performed this period. A facility management plan is about how the facility is run, covering delivery model, service levels, team, space and budget.
The plan does contain a performance-measures table, which is where the two touch. Keep strategy and budget in the plan and keep the calculations here, rather than maintaining two versions of the same measure set.
How should RAG thresholds be set?
In numbers, in the file, before the first report is produced. The template suggests green within a small percentage of target and amber within a wider band, with red beyond it, and asks you to write the actual figures down.
The reason is simple. A status decided after the result is known is not a measurement, and everybody reading the report can tell.
Should every measure have an action against it?
No. The file deliberately restricts actions to the measures that are amber or red, with an owner and a date on each.
A scorecard where every line carries an action produces a list nobody works through, which quickly means none of the actions are real, including the ones that mattered.
How do you set targets that people take seriously?
Record how the target was set, which is a field in the template: benchmark, contract, previous performance or aspiration. That one note changes how a miss is read.
A target taken from a benchmark for a different portfolio is the common failure. It gets missed every period, and after two quarters of unexplained red the whole scorecard loses credibility rather than just that measure.
Can this be used to compare several sites?
That is largely what the definitions are for, and it works as long as every site uses this set of rules rather than its own. The Trend tab holds one row per period for a single scorecard.
The practical limit is manual comparison. Once you are asking which of several sites has been red twice running, you are opening a workbook per site, which is the point at which the spreadsheet stops being the right tool.
In one paragraph
A facility management KPI set is only as good as its definitions. Write the calculation for every measure out in words, including what does not count, because exclusions are where two honest people produce different numbers for the same measure. Name a source, an owner and a frequency for each. Agree the RAG thresholds and record how targets were set before the first report, then raise actions only on the lines that are amber or red.
The scorecard is the floor, not the ceiling
Define the measures once, report them consistently and keep the action list short. Once every figure is still being collected by hand, or you are comparing sites by opening a workbook each, the spreadsheet has done its job.
